Prediction of Lanzhou urban passenger flow based on machine learning

Abstract

Due to the continuous occurrence of black swan events, the large number of tourists and the uncertainty have caused certain challenges to the tourism industry, so it is very useful to accurately predict the tourist flow in tourism layout and management. For this reason, this paper takes the time series of Lanzhou tourism flow in 2019 as the research object and analyzes its trend. To improve the prediction accuracy, we propose a model that combines empirical mode decomposition (EMD), particle swarm optimization (PSO), and gated recurrent unit neural network (GRU), and combines it with several classical Compare time series forecasting models. Experiments finally show that this method can well reduce the hysteresis of GRU model prediction, and can quickly find the optimal parameters of the neural network, with more accurate prediction results.